Gaps in the National Policy Framework on AI Leave State Regulations in Limbo, Report Claims

The analysis lists five major areas of concern on which the framework is mostly silent, leaving state attempts to address them in a kind of regulatory limbo: algorithmic accountability; data privacy beyond children; transparency requirements; enforcement mechanisms; and workforce displacement.

Regarding algorithmic accountability, the framework says nothing about any requirements for AI systems to be audited for discriminatory outputs, Shark notes. Yet it threatens to preempt state laws that impose “AI-specific” requirements, such as those in New York and Colorado that mandate bias assessments for AI systems used in sensitive decisions such as hiring, lending, housing or benefit eligibility while establishing no federal equivalent.

While the National Framework includes guidelines for federal regulation to provide data privacy protection for children it is silent on privacy protections for adults, the analysis points out. That leaves it at odds with state laws such as the California Consumer Privacy Act and those in other states that impose privacy regulations on organizations that process large amounts of sensitive personal data, which again could afoul of federal preemption of AI-specific state-level mandates.

Transparency and explainability requirements is another area where the framework is silent. It says nothing, for instance, about ensuring transparency on public agencies’ use of AI in consequential decisions such as tax assessments and permit approvals, or consumers’ right to challenge those decisions, which Shark describes as “fundamental to due process.”

Related: SEC and CFTC Release First-Ever Crypto Classification Framework 

While the framework pays lip service to expanding AI training for workers, the analysis notes it explicitly favors “non-regulatory methods” and calls for studying AI’s impact on jobs and employment trends. That leaves local officials grappling with communities facing real job displacement risks with little to hang their hats on.

Finally, per Shark, the framework contains no compliance deadlines, nor directs any federal agency to take specific actions with respect to its own recommendations, making it more aspirational than operational.

The effects of that silence are already being felt. Even before the framework was released, the analysis notes, the Trump administration showed itself to be hostile toward state efforts to regulate AI systems and companies, including creating a task force in the Justice Department to target state laws that conflict with federal policy. States weighing new AI transparency mandates or algorithmic accountability requirements, the analysis says, must now factor in the legal and financial risk of being targeted.

The White House framework also falls short for AI companies hoping for a clear federal standard that would override the dense thicket of state regulations, according to the analysis. “The framework’s own legislative path is uncertain, with Democratic opposition and intra-Republican tension over states’ rights complicating the math,” it notes. “California, Colorado and New York compliance regimes remain in full force,” and AI vendors operating across multiple states “will continue navigating a multilayered legal environment.”

The bottom line, Shark concludes, is the framework is long on what the administration would like to see happen and not happen in the realm of AI, but short on how to get there.